358 Mr. A. 8. Wood ward on the 



supposed it might jjortain to a so-called Pycnodont Ganoid ■**■ ; 

 and in Owen's * Palaeontology ' (edit. 2, 1861, p. 174) 

 Pisodus is also doubtfully quoted as a " Ganoid " of uncertain 

 position. 



It now appears from a nearly complete skull in the British 

 Museum that the problematical fossil in question is the para- 

 sphenoid dentition of a fish remarkably similar in cranial 

 characters to the recent Clupeoid Albula. The fact has 

 already been incidentally mentioned in a record of the disco- 

 very of Pisodus in the Middle Eocene of Belgium t ; and it 

 only remains to justify, by a detailed description and figures, 

 the recognition of an Alhula-\\\ie. fish at so remote a period as 

 that of the Lower Eocene. Dr. Shufeldt's admirable 

 description of the skull of the recent Albula vulj^es fortunately 

 suffices for requisite comparison |. 



The fossil skull in question is imperfectly preserved in 

 several respects, though exhibiting most of the principal 

 characters. It is shown of one half the natural size, from 

 four aspects, in the accompanying Plate XVII. The rostral 

 region is wanting, the projecting lateral portions are broken 

 away, and the basioccipital and basisphenoid bones are in 

 great part removed by fracture. The general form of the 

 cranium, however, is recognizable, and the great expansion of 

 the parasphenoid, with its tritoral dentition, is especially well 

 disj)layed. The cranial roof (fig. 1) is flattened and termi- 

 nates abruptly behind in a straight occipital border. The 

 occipital face (fig. 2) is nearly vertical ; and the basioccipital 

 axis is approximately parallel with the cranial roof, meeting 

 the dentigerous portion of the parasphenoid in a sharp angu- 

 lation (fig. 4). 



On the occipital face (fig. 2) the two exoccipital elements 

 {ex.occ.) are observed immediately above the foramen 

 magnum {f. rn.), and seem to have been originally apposed 

 in the median line except at their ujiper extremity. These 

 bones are deeper than broad, so far as preserved, quadrilateral, 

 and with a slightly concave outer surface. Their superior 

 internal angles are truncated for the reception of the lower end 

 of the large triangular supraoccipital {s.occ), which is flat- 

 tened, but shows remains of a T-sIiaped, backwardly directed, 

 mesial crest (c) in its upper two thirds. 



• L. Apassiz, Poias. Foss. vol. ii. pt. ii. p. 247 (1844). 



+ Smith Woodward, " Notes on some Fish-remains from the Lower 

 Tertiary and Ujtper Cretaceous of Belgium," Geol. Mag. [3] vol. viii. 

 p. 108 (18'Jl). 



t K. W. Shufeldt, "The Osteology oi' A inia vulva A:c.," Ann, Kep. 

 I.S. Fiish Coiiimis'jion, 188.''>, pj). Hl-O", ]il.s. xii., xiii. 



